Pointing Accuracy

Ang. dist.

-2 -1 0 1

y to cog [mrad]

-2 -1

Abbildung 4.2: (Left) The structure design of the H.E.S.S. telescope. (Right) On-axis intensity distribution of a star on the closed camera lid after ali-gnment of the mirrors. The hexagonal border is the size of a pixel in the camera. Figure taken from [50].

4.4 Pointing Accuracy

The accuracy of the pointing of a telescope depends on the mirror alignment accuracy, the tracking accuracy, and the precision of corrections for misali-gnments and deformations of the structure of a telescope.

The mirror alignment procedure is based on viewing the image of a star, which is reflected by facet-mirrors onto the closed lid of the telescope’s ca-mera, [50]. The reflected image positions (spots) are recorded by a lid-CCD camera, which is mounted at the center of the telescope’s dish (Figure 4.3).

Then each of the facet-mirrors are moved one by one, changing the locati-on of the corresplocati-onding spot locati-on the lid; the displacement is recorded every time. With the help of this recorded data, the spot can be focused into a single position at the center of the camera lid. This procedure guaranties that a point-like source at infinite distance to be imaged on the focal pla-ne of the telescope. The point spread function (PSF) shown in Figure 4.2 (right) is the on-axis intensity distribution of a star on the camera lid after alignment. It can be seen that the distribution is symmetrical and the width is smaller than the size of a PMT pixel, which is about 0.16^◦. On the opti-cal axis, an RMS-width of ∼48” is measured. When this accuracy of aligned mirrors is combined with the tracking precision (≈30”), the telescope points

Abbildung 4.3: The position of lid-CCD and sky-CCD on the dish indicated by white circles. The lid-CCD is mounted in the center of the telescope dish, and the sky-CCD can be seen on the left of the lid-CCD.

to a source with a combined accuracy of ∼60”. Detailed information on the mirror-alignment procedure can be found in [26], [51]

The mis-pointing of the telescopes due to the mechanical imperfections should be corrected in order to achieve an arc-second level of resolution. The mechanical imperfections are the reproducible errors such as bending of the structure under gravity, as well as the irreproducible effects like bending due to wind. To predict the reproducible errors positions of a sample of bright stars are recorded (while the lid of the camera is closed) by the lid-CCD as a function of azimuth and altitude. A detailed mechanical model is fit to this data so that given a pointing of the telescope the model returns the expec-ted mis-pointing, [69]. Without any model corrections, the raw mechanical pointing accuracy is about 60” (two-dimensional RMS), after the correction, a pointing precision of 20” RMS is achieved. This accuracy can be impro-ved by other techniques using the sky-CCD (shown in Figure 4.3). These techniques are still under study, [70].

4.5 Camera

The camera of a H.E.S.S. telescope consists of 960 pixels (PMTs) each having a diameter of 0.16^◦, which results in an overall field of view of 5^◦. All the signal processing, triggering, and digitization processes run within the camera body.

Abbildung 4.4: After mirror alignment a second step correction due to the mis-pointing of the telescope is done. (Left) The spots of the stars on the camera plane before applying any pointing-model have a RMS of about 60”.

After applying the pointing corrections a pointing precision of less than 20”

is obtained. Figure taken from [70].

The photo-multiplier tubes (PMTs) of type Photonis XP2960 consist of a photo cathode enclosed with µ-metal shielding, which prevents the photo-electrons to be deflected by the Earth’s magnetic field and plastic casing. The material, of which the PMT’s window is made affects the sensitivity of the PMT. In order to be able to detect Cherenkov light, the transparent range has to be above 250 nm. The PMT’s used in the H.E.S.S. experiment have windows made of borosilicate glass. A PMT acts as a combination of a sim-ple photo-cell with a high-gain amplifier. The gain for each PMT is 2×10⁵. Operating voltages for the PMTs are supplied by DC-DC converters integra-ted into each PMT base with active stabilization for the last four dynodes for best linearity, [97]. The quantum efficiency, Q_{ef f} can be expressed as the number of electrons emitted by the cathode divided by the number of photons hitting the PMT’s window. The quantum efficiency depends on the material that the window and the cathode is made of. The maximum Q_{ef f} is around 25% for the wavelength range from 390 to 420 nm, [101].

The collection of light that is reflected from the telescope’s mirrors onto the camera is improved by Winston cones, which are mounted on top of the pixels in the camera. These light-guiding cones are used to salvage photons, which would otherwise be lost in the space between pixels. They are also used to select light that comes from the mirrors of the telescope and shield against large impact angle scattering light etc..

16 pixels are grouped in a drawer unit. There are 60 drawers in each

Abbildung 4.5: The H.E.S.S. Camera with the lid opened.

camera. The drawers can be inserted from the front side of the camera body and have connectors to power, readout bus, and trigger bus at the backside.

Apart from the PMTs, each drawer contains two acquisition cards, each of which is connected to 8 PMTs. The incoming signals from the PMTs are measured across a resistor,R_{P M}, and amplified into the acquisition channels.

In order to observe the showers within an energy range from 40 GeV to 20 TeV or higher, a dynamical range of 1 - 2000 photoelectrons (ph.e.) is required. With only one channel per PMT it is not possible to cover this range. Therefore, for each pixel there are two acquisition channels that have different gains: the high-gain (HG) channel is used to detect signal charges up to 200 ph.e., and the low-gain (LG) channel covers the range from 10 to 1600 ph.e.. For the analysis, the linear range (Figure 4.6) of both of the channels is used.

Apart from the acquisition channels, each acquisition card has a trigger channel and four Analog Ring Samplers (ARS)s, each of which contain 128 capacitor cells. The ARS plays an important role in the camera’s readout process. Analog signals arriving from the (HG and LG) acquisition channels of a PMT are sampled in the ARS. This is done by storing the analog voltage levels in a ring buffer of 128 capacitor cells with a rate of 1 GHz (i.e. each cell is stored within 1 ns). Sampling of analog signals continues until a camera triggersignal arrives. Usually it takes 70 ns that a trigger signal is formed (see also Section 4.6). In each ARS the width of the read-out window, where the Cherenkov signal is expected, is normally 16 cells long. Only the analog signal from this read-out window is digitized by an Analog to Digital Converter (ADC)with a specific conversion factor of 1.22 mV per ADC-counts ([88]) and

Abbildung 4.6: (Left) Linearity of the high-gain (HG) channel and (right) linearity of the low-gain (LG) channel. Figure taken from [88].

sent to a Field Programmable Gate Array (FPGA) that allows two different modes of recording: the charge-mode and thesample-mode. In sample-mode ADC-values stored in each cell of the read-out window are recorded. In the charge-mode the ADC-values of all cells of the read-out window are summed and recorded. The sample-mode is used in calibration (Section 4.9) of the data, the charge-mode for data-taking, [164].

4.6 Trigger

In the H.E.S.S. experiment simultaneous observation of an air-shower with multiple telescopes is carried out at the hardware level by the central trigger system (CTS). The CTS consists of acentral trigger unitplaced in the control building of the H.E.S.S. array and interface modules located in each camera.

The communication between the CTS and interface modules is achieved by an optical fiber link.

The trigger-process consists of two parts: The camera trigger and the central trigger. The camera triggerresults from a multiplicity trigger within overlapping trigger sectors(Figure 4.7), each having 64 pixels. In a pixel, the PMT comparator checks if the number of ph.e. in the signal exceeds a given pixel threshold, p, and creates a trigger pulse. The length of the pulse reflects the time the input signal exceeds the threshold. Since typical noise signals rarely exceed the threshold pand result in short trigger signals, the effective resolving time of the pixel coincidence is in the 1.3 to 2 ns range providing

Abbildung 4.7: Each of the H.E.S.S. cameras is divided into 38 overlapping trigger sectors.

a high suppression of random coincidences, [97]. The camera trigger occurs by the coincidence of the number of pixels, which have signals above the threshold p, in a trigger sector exceeding an adjustable sector threshold, q (usually from 3 to 5 pixels). The time needed to build a trigger signal is 70 ns. Following a camera trigger signal the readout of the camera starts. At the same time, the trigger information of the camera is sent to the CTS.

In the CTS if several telescope triggers coincide within a time window of 80 ns, a central trigger signal is send. This signal initiates the readout from the data of the whole array. The CTS assigns an absolute time-stamp to each system-event, which is provided by a GPS clock in the central trigger unit, [65].